J. Agric. Food Ch8m. 1990, 38,397-402
397
Variation in the Extractability of Esterified p-Coumaric and Ferulic Acids from Forage Cell Walls Hans-Joachim G. Jung*pt and Snake C. Shalita-Jones' U.S. Dairy Forage Research Center, USDA-ARS, and Department of Animal Science, University of Minnesota, St. Paul, Minnesota 55108
Extraction conditions that may affect yield of esterified phenolics from forage cell walls were examined. Esterified p-coumaric and ferulic acids were extracted from alfalfa, smooth bromegrass, and switchgrass samples with various concentrations of NaOH. Forage fiber samples were prepared prior to alkaline extraction by enzymatic removal of starch and alcohol extraction (AIR), extraction with neutral detergent (NDF), or extraction with neutral detergent followed by ball-milling to reduce particle size (BM-NDF). The forage species had different phenolic acid contents, with alfalfa having the lowest and switchgrass the greatest. An interaction was observed for sample preparation method and alkali concentration. The AIR and NDF gave maximal yields of p-coumaric and ferulic acids from alfalfa with use of 2 M NaOH for extraction, whereas the yield from grasses peaked a t 1 M. Recovery of p-coumaric and ferulic acids was greater from the AIR preparation than NDF for all forages. NaOH a t 4 M concentration was needed for maximal yield from BM-NDF of grasses, and 6 M concentration was best for alfalfa BM-NDF. Immature and mature growth stage samples of forages behaved differently in response to extraction conditions. The molar ratio of p-coumaric to ferulic acids extracted from forage samples changed in response to different preparation conditions and alkali concentrations. Recommendations are given for maximizing yields of extractable phenolic acid esters from forage cell walls.
p-Coumaric and ferulic acids have been found to be esterified to the cell walls of grass ahd legume species (Jung et al., 1983a). Concentrations of these esterified phenolic acids are greater in grasses than legumes, but the distribution of phenolic acids is different between anatomical fractions of these forages. In corn (Zea mays) herbage, stem concentrations of p-coumaric and ferulic acids were greater than those in leaves (Hartley and Haverkamp, 1984); however, the opposite result was observed in alfalfa (Medicago sativa) (Bittner, 1984). Physiological maturity of forages is known to alter the concentrations of esterified p-coumaric and ferulic acids (Jung et al., 1983b; Burritt et al., 1984). These esterified phenolic acids have been implicated as inhibitors of ruminal fermentation of cell wall polysaccharides based on negative correlations observed between in vitro digestibility of grasses and the p-coumaric to ferulic acid ratio of the forage (Hartley, 1972; Burritt et al., 1984). As the ratio of these phenolics increased, digestibility was observed to decline. Concentration of p-coumaric acid, but not ferulic acid, also was negatively correlated with ruminal digestibility of grasses (Burritt et al., 1984). The ability of these cinnamic acids to inhibit ruminal fermentation of polysaccharides has been corroborated with model systems using free and ester-linked p-coumaric and ferulic acids (Akin, 1982; Jung and Sahlu, 1986a). Virtually all the data on esterified cinnamic acid concentrations in forage cell walls were derived by extraction of neutral detergent fiber (Goering and Van Soest, 1970) with 1 M NaOH. The purpose of this study was to evaluate the effect of sample preparation method to remove non cell wall components, sample particle size, and alkali concentration on yield of p-coumaric and ferulic acids from various forages.
USDA-ARS and University of Minnesota. University of Minnesota.
MATERIALS AND METHODS Forage Sample Preparation. Alfalfa, smooth bromegrass (Bromus inermis), and switchgrass (Panicum uirgatum) were
chosen as forage sources as they represent the three major taxa of forages (C, legume, C, grass, and C, grass, respectively) normally fed to ruminants and are known to differ in cell wall chemistry (Akin et al., 1984). A single sample of each forage species was harvested at a vegetative, immature growth stage and again after development of mature seed. Total herbage was collected by clipping the plants 2 cm above ground level. Forage samples were frozen, lyophilized, and ground through a 1-mm screen in a cyclone mill (Tecator AB, Hoganas, Sweden). Non cell wall forage components were removed by extraction with neutral detergent (Goeringand Van Soest, W O ) , minus sodium sulfite, or by treatment with amylase to remove starch and extraction with 80% ethanol to remove alcohol-solublephenolic glycosides. The neutral detergent fiber (NDF) was prepared by refluxing the forage samples with neutral detergent (100 mL/g sample) for 60 min. The NDF was then filtered and washed 4 x with large volumes of hot water and 2X with acetone. The NDF samples were allowed to air-dry in a hood. The starchfree alcohol-insolubleresidue (AIR) was prepared according to Theander and Westerlund (1986). Acetate buffer (25 mL, 0.1 M, pH 5.0) and 0.5 mL of heat-stable a-amylase (Sigma Chemical Co., St. Louis, MO) were added per gram of forage sample and heated at 90 "C for 60 min. After the mixture was cooled to 50 "C, 1mL of amyloglucosidase (Sigma)was added per gram of sample and the resultant mixture heated for 3 h at 60 "C. Sufficient 95% ethanol was added to achieve a final concentration of 80% ethanol, and the sample was held at 4 "C overnight. The AIR was recovered by centrifugation, washed 2X with 80% ethanol and once with acetone, and allowed to airdry under a hood. A third sample preparation treatment with a smaller particle size was prepared by ball-milling NDF (BMNDF) for 1 min at 4 "C in a shaker mill (Pica Blender Mill Model 2601; Hankison Co., Canonsburg, PA). Dry matter was determined at 100 "C, and organic matter was calculated after ashing at 450 "C for 4 h. Extraction of Esterified Cinnamic Acids. Prepared forage samples were extracted with 0.5, 1, 2, 4, or 6 M NaOH by a modification of the procedure of Jung et al. (1983a). Samples (200 mg) were placed in 25 X 150 mm glass culture tubes
0021-856119011438-0397$02.50/0 0 1990 American Chemical Society
380
Jung and ShalitaJones
J. Agric. Food Chem., Vol. 38, No. 2 , 1990
Table I. Recovery of Organic Matter from Forages after SamDle PreDaration recovery,' mg/g OM AIR NDF forage
immature
alfalfa smooth bromegrass switchgrass
766 879
a
940
mature 1007 991 1008
immature
308 552
647
mature 667 702 768
Results are expressed relative to the organic matter content of
the original sample.
with Teflon-line caps. The appropriate anaerobic NaOH solution (10 mL) was added to each tube under a N, stream. Samples were extracted at 39 "C for 24 h in the dark with occasional shaking. Samples were subsequently filtered and washed with 5 mL of water. The combined filtrate and wash was acidified to pH 2.6 with concentrated phosphoric acid. The acidified sample solution was loaded on a c,, solid-phaseextraction column (Supelco, Inc., Bellefonte, PA), the column was washed with 2 mL of NaOH-phosphoric acid solution (pH 2.6), and the cinnamic acids were eluted with two 2.5-mL washes of 50% methanol. The eluted samples were brought to a final volume of 10 mL, filtered through a 0.2-pm filter, and stored at -20 "C until analyzed. All extractions were performed in duplicate. Cinnamic Acid Analysis. Identification and quantification of phenolic acids was done by LC (Jung et al., 1983a). A Gilson gradient autoanalytical system (Gilson Medical Electronics, Inc., Middleton, WI) with a programmable dual-wavelength UV detector (Gilson Model 116) was utilized. Samples (20 FL) were injected and separated by isocratic elution through a Spherisorb-ODs,CIS, 5-pm column (Supelco). The solvent (97.7% water-O.3%glacial acetic acid-2% butanol) was pumped at 3 mL/min. Column temperature was maintained at 50 "C. The chromatographicseparation was developed to detect 10 different phenolic acids and aldehydes, and both 254 and 320 nm were monitored simultaneously. p-Coumaric and ferulic acids were quantified at 320 nm, and the ratio of absorbances at 254 vs 320 nm was used to check compound identification. Statistical Analysis. The forage species were known to differ greatly in phenolic acid content, and therefore, the data were analyzed separately for each forage. The statistical model employed WBS a 3 X 2 X 5 factorial of sample preparation method, forage maturity,and alkali concentration. Individual means were tested by the F-protected least significant difference method (Steel and Torrie, 1960). RESULTS The recovery of forage organic matter in the AIR and NDF prepared samples is shown in Table I. The recovery of organic matter from BM-NDF was assumed to be the same as NDF. The virtually complete recovery of organic matter by the AIR preparation suggests that residual protein and lipid material remain in the AIR, as observed previously (Selvendran, 1975). Treatment of cycellulose and oatspelts xylan (Sigma) with the AIR procedure resulted in organic matter recoveries of 104-117 %. This suggests that partial acetylation of the cell wall polysaccharides occurs due to heating in acetate buffer, which would add organic matter to the samples. As expected, NDF content of the legume was least and the C, grass, switchgrass, contained the highest concentration of NDF. Forage NDF content was also increased with physiological maturity of the sample. Similar general trends were seen for the AIR. Although not quantified, the leaf to stem ratio of the forages declined with advanced maturity. The forage samples provided a broad spectrum of both forage types and fiber content. Because organic matter recovery relative to the original plant material was different among the sample preparation methods, the yields of phenolic acids reported were corrected
to the initial organic matter content of the forage samples. Degradation of p-coumaric and ferulic acids at the different NaOH concentrations was determined by treating a mixture of pure phenolic standards, mixed with glucose as a carrier to facilitate handling, with each alkali concentration in the same manner as the forage samples. Phenolic standards were not mixed with forage samples to avoid matrix interactions that might limit recovery of the free acids. Since free phenolic acids are present in extremely small amounts, such matrix-phenolic interactions were not considered valid. Degradability of phenolic acids in alkali was relatively consistent. The recovery of p-coumaric acid averaged 74.6 f 0.7% and did not differ ( P > 0.05) among the NaOH concentrations. Ferulic acid recovery was greater ( P < 0.05) at 0.5 M NaOH (79.7%) than at the other concentrations (75.0 f 0.8%),which were not different ( P > 0.05). The data on yield of p-coumaric and ferulic acids from forages were corrected for their recovery a t each NaOH concentration. Statistical analysis of the esterified phenolic acid yield data from the different sample preparation methods, alkali concentrations, and forage maturity stages revealed a very complex situation. For every forage species the three main effects (alkali concentration, sample preparation method, and forage maturity) were always significant ( P < 0.05) as were the two-way interactions, except for alkali concentration X forage maturity interaction, which was significant in half the cases (Table 11). Ferulic acid yield even exhibited a significant three-way interaction of preparation method X"alkali concentrations X forage maturity for both alfalfa and smooth bromegrass. The alkali concentration X forage maturity interactions resulted from a smaller difference in phenolic yield, but still significant, between maturity stages for the AIR preparation method than seen for the other two procedures. The observed three-way interactions were more difficult to interpret but appeared to be related to small differences as to which alkali concentration provided maximal phenolic acid yields relative to forage maturity in each of the preparation methods. Further discussion will be limited to the greater differences associated with main effects and the alkali concentration interactions with preparation method and forage maturity. The yields of p-coumaric and ferulic acids from the different sample preparation procedures, after extraction by various NaOH concentrations, are shown in Figures 1-3. Mean values, across maturity stages, are given because differences were consistent for both immature and mature forage samples. For alfalfa, alkali concentrations had relatively little effect on release of phenolic acids from AIR or NDF; however, reduction of particle size resulted in greater yields with increasing alkali strength from BM-NDF (Figure 1). The AIR sample preparation method resulted in consistently greater (P 0.05) yields of both p-coumaric and ferulic acids than obtained from NDF. p-Coumaric acid yield from alfalfa was greater ( P < 0.05) from NDF than BM-NDF at 0.5,1, and 2 M NaOH concentrations, but similar at higher alkali levels (Figure la). Release of ferulic acid from alfalfa was greater ( P < 0.05) from NDF than the BM-NDF preparation at all NaOH concentrations (Figure lb). A different pattern of response was seen for smooth bromegrass. For both AIR and NDF, release of phenolic acids increased as alkali strength changed from 0.5 to 1 M NaOH, but yields declined with further increases in NaOH concentration (Figure 2). In contrast, p-coumaric and ferulic acid yields increased from BM-NDF
J. Agric. Food Chem., Vol. 38, No. 2, 1990 399
Phenolic Ester Extraction from Forage Cell Walls
Table 11. Summary of Statistical Analysis for the Extractability of Pbenolic Acids due to the Variables Studied. main effects interaction NaOH concn prep method forage maturity alfalfa smooth bromegrass
p-coumaric
switchgrass
ferulic p-coumaric ferulic
Key:
-
(4 ** ** ** **
phenolic acid p-coumaric ferulic
forage species
(B) **
(C)
AXB
AxC
** * ** ** ** **
** ** ** ** ** **
** ** ** ** ** **
** ** **
**
** **
**
BxC
NS
*
NS NS NS
NS
**
* **
NS
NS
*, **, significant effect at P < 0.05 and 0.01, respectively; NS, nonsignificant (P> 0.05).
020r
3s0
i
1.50
0.00 0
I
~
0.25
I
2
3
4
5
6
7
a
..
& , .,
h
AxBxC
*
1 I
1.00 0
r 4'00
I
2
1
1
3
4
5
7
6
b
3.00
2
/ 010L
2.50
**/------+-----4
------c
*------
k 3
&
L
005c
~3-
I
I
000 I 0
1.50 1
2
3
4
5
6
I
7
Concentration of NaOH (mol/l)
Figure 1. Yields of p-coumaric (a) and ferulic (b) acids from alfalfa samples extracted with alkali. Values are averages of AIR (m),NDF (e),and BM-NDF (0) from immature and mature forage samples. Standard errors of the mean were 0.006 and 0.005 mg/g of OM for p-coumaric and ferulic acids, respectively. smooth bromegrass as alkali strength increased from 0.5 to 4 M, which was followed by a decline a t 6 M NaOH. Yields of phenolic acids were greater ( P C 0.05) from the AIR than the NDF preparation a t all NaOH concentrations, as observed for alfalfa. Unlike alfalfa, extraction of p-coumaric and ferulic acids from smooth bromegrass was greater ( P C 0.05) from the BM-NDF than found for the NDF at 2, 4, and 6 M NaOH. p-Coumaric acid yield also was greater ( P < 0.05) a t 0.5 M NaOH from BM-NDF than NDF (Figure 2a). The response of switchgrass to sample preparation method and NaOH concentration was more similar to smooth bromegrass than alfalfa. Maximum yields of both p-coumaric and ferulic acids from switchgrass AIR and NDF preparations were found with 1M NaOH, and yields declined a t higher alkali concentrations (Figure 3). As in smooth bromegrass, phenolic acid release from switchgrass BM-NDF peaked a t 4 M NaOH. Yields of both phenolic acids were greater ( P C 0.05) from AIR than NDF of switchgrass. Unlike alfalfa or smooth bromegrass, p-coumaric acid release from BM-NDF was con-
sistently greater (P C 0.05) than from NDF in the case of switchgrass. Degree of forage physiological maturity appeared to influence the content of phenolic acids in different sample preparations. For all forages and sample preparation methods, yields of p-coumaric and ferulic acids were different ( P C 0.05) between immature and mature forage material when averaged across alkali concentrations (Table 111). Release of p-coumaric was always greater ( P C 0.05) from mature forage than immature samples. In contrast, ferulic acid yield from mature alfalfa was greater ( P < 0.05) then from immature, but the opposite was seen for both smooth bromegrass and switchgrass. Sample preparation method also influenced yield of phenolic acids. Both p-coumaric and ferulic acid yields from immature and mature forage samples were greater ( P C 0.05) from the AIR than from NDF (Table 111). Reduction of particle size resulted in greater ( P < 0.05) phenolic acid yields from BM-NDF than seen with the NDF preparation for the grass samples, but the comparative
Jung and $hathaJones
J. Agric. Food Chem., Vol. 38. No. 2, 1990 7l
4
a
Figure 3. Yields of p-coumaric (a) and ferulic (b) acids from
switchgrass samples extracted with alkali. Values are averages of AIR (m),NDF (e),and BM-NDF (0) from immature and mature forage samples. Standard errors of the mean were 0.17 and 0.09 mg/g of OM for p-coumaric and ferulic acids, respectively. yields from alfalfa were variable between the sample preparation methods. Not only are yield of p-coumaric and ferulic acids different among sample preparations and alkali strengths, but the relative proportions of the phenolic products also vary. Table IV illustrates the differences among extraction conditions for p-coumaric to ferulic acid molar ratio when the highest yielding alkali concentration for each forage and preparation method was used. Except for BMNDF of alfalfa, all immature and mature forage samples treated the same were different ( P < 0.05) in phenolic acid ratio, with proportionately more p-coumaric acid in the mature samples. Within a forage sample, the AIR provided the lowest (P < 0.05) ratio or overlapped with that for BM-NDF. DISCUSSION The differences noted in p-coumaric and ferulic acid contents of alfalfa compared to the grass species have been observed previously (Jung et al., 1983a,b). Changes in yield of phenolic acids due to physiological maturity of the forage also agreed with previous work (Burritt et al., 1984). Both ferulic and p-coumaric acids are esterified to hemicellulosic components and core lignin in the cell wall, but ferulic acid is predominantly associated with the polysaccharide fraction while p-coumaric acid is primarily attached to core lignin (Atsushi et al., 1984; Azuma et al., 1985). The increase in the p-coumaric to ferulic acid ratio with forage maturation (Table 111) can be explained by the reduction in hemicellulose synthesis or concentration and increase of core lignin in the cell wall
that occurs during maturation (Jung and Vogel, 1986b). The yield differences of phenolic acids from NDF compared to AIR (30-40% less from NDF) were striking. It appears that the neutral detergent procedure results in loss of forage cell wall material that contains appreciable amounts of esterified phenolic acids. It has been shown that forage cell wall content estimated by NDF is up to 25% less than the sum of cell wall neutral sugars, uronic acids, and Klason lignin (Theander and Aman, 1980). The loss of cell wall by the NDF procedure was greatest from legumes, and across forage species most of the loss was from uronic acids, arabinose, rhamnose, and galactose fractions (Theander and Aman, 1980). Fry (1982,1983) has shown that ferulic acid, and some p-coumaric acid, is esterified to arabinose and galactose units of spinach (Spinucia oleracea) pectic substances. Feruloylated pectins also have been reported from sugar beet (Beta saccharifera) pulp (Rombouts and Thibault, 1986). Although the pectins in forages have not been well studied, it appears that loss of pectic substances from the plant cell wall by neutral detergent extraction could account for some of the difference in yield of phenolic acids, especially ferulic acid (Table 111), between the AIR and NDF preparations. It has previously been found that intact forage samples contain more phenolic acids than the NDF fraction (Jung et al., 1983a). From these data it was concluded that 1740% of the esterified p-coumaric and ferulic acids in forages were in the cell soluble fraction rather than the cell wall. Cherney et al. (1989) have recently reported similar results for eight grass and five legume species. However, the loss of cell wall polysaccharides containing esterified phenolic components by neutral detergent extraction would indicate that NDF-based estimates of cell soluble phenolics are too high. This leaves the question of whether the AIR preparation of Theander and Westerlund (1986) solubilizes all of the non cell wall bound esterified phenolics. Mosihuzzaman et al. (1988) found 36 and 6% of total p-coumaric and ferulic acids, respectively, were in the cell-soluble fraction using 80% ethanol extraction in one species of jute (Corchorus capsularis), but
